Dispersant Effects in the Selective Reaction of Aryl Diazonium Salts with Single-Walled Carbon Nanotubes in Aqueous Solution Adam J. Blanch, Claire E. Lenehan, and Jamie S. Quinton* Flinders Centre for Nanoscale Science & Technology, School of Chemical and Physical Sciences, Flinders University, GPO Box 2100, Adelaide SA 5001, Australia
bS Supporting Information ABSTRACT:
Current methods of synthesis for carbon nanotubes (CNTs) usually produce heterogeneous mixtures of diﬀerent nanotube diameters and thus a mixture of electronic properties. Consequently, many techniques to sort nanotubes according to their electronic type have been devised. One such method involves the chemical reaction of CNTs with aryl diazonium salts. Here we examine the reactions of electric arc produced CNTs (dispersed by a variety of surfactants and polymers in aqueous solution) with 4-bromo-, 4-nitro-, and 4-carboxybenzenediazonium tetraﬂuoroborate salts in order to ﬁnd conditions for maximum selectivity. Reactions were monitored through the semiconducting S22 and metallic M11 transitions in the UVvisNIR absorbance spectra of the nanotube dispersions. Selectivity was observed to depend heavily on the type of surfactant, the type of diazonium salt and its concentration, the reaction temperature, and the solution pH. Additionally, the surfactant concentration was found to exert a signiﬁcant inﬂuence as the dediazoniation product yields are aﬀected by this parameter. For certain combinations of surfactant and diazonium salt the selectivity is markedly improved, particularly in dispersions of nonionic surfactants Pluronic F-127 and Brij S-100, which are similar in structure. Smaller diameter HiPCO nanotubes were better functionalized in dispersions of Triton X-405. The greater selectivity aﬀorded by these poly(ethylene oxide) containing polymers is postulated to arise from electron donation provided by their ether oxygens. The ionic surfactant sodium dodecyl sulfate was found to display unique behavior in that semiconducting nanotubes were preferentially functionalized at natural pH, likely due charge localization interactions with the surfactant.
1. INTRODUCTION The inhomogeneous nature of carbon nanotubes (CNTs) in their as-produced form presents a considerable barrier to their implementation in a diverse range of applications. CNTs in the raw “soot” provided by most bulk synthesis methods are generally composed of a broad range of lengths, diameters, and electronic properties, while also being strongly bound to each other in ropes and bundles through van der Waals forces.1,2 A signiﬁcant amount of carbonaceous and metallic impurities is also present in most commercially available nanotube products, which further complicates processing of these materials.3 Ultimately, control of CNT growth should aﬀord nanotubes of uniform properties in the foreseeable future; however, in the interim methods of purifying and sorting CNTs have garnered much interest, and many techniques have been investigated for r 2011 American Chemical Society
this purpose.2,4,5 Recently, separation through the use of density gradient ultracentrifugation6 has become increasingly popular, owing to its noncovalent nature and ability to distinguish between CNTs of diﬀerent diameter and even chiral handedness.7,8 However, this method produces relatively small quantities of puriﬁed material in a single run; thus, alternative methods may be more suited for producing large quantities of material that are not necessarily enriched in a single nanotube species. Selective dispersion using speciﬁc DNA sequences9 also shows promise, though the synthesis of oligonucleotides increases the cost of the process and the strength of the binding interaction Received: August 24, 2011 Revised: November 17, 2011 Published: December 06, 2011 1709
dx.doi.org/10.1021/jp208191c | J. Phys. Chem. C 2012, 116, 1709–1723
The Journal of Physical Chemistry C makes the dispersant diﬃcult to remove. Gel chromatography is another separation technique that has demonstrated excellent potential for large-scale sorting of CNTs; however, nanotubes with diameters 1.4 nm and above are currently unable to be separated using this method.10 Thus, despite the considerable progress that has been made in the ﬁeld of nanotube separation, there is room for advancement. The reaction of CNTs with diazonium salts11 has been used extensively as an approach for covalent functionalization of the CNT sidewall thanks to its simplicity, scalability, and relatively low cost, with some salts being commercially available. 12 Although the reaction has been shown to be selective toward CNTs that are metallic in nature over their semiconducting counterparts,13 only a small number of studies have been performed using diazonium salts for separation of CNTs by electronic type.1419 This is probably due to the limited success reported so far using this method, though relatively few combinations of surfactant and functional constituents of the diazonium salt have been assessed for their viability.12,1722 While the diazonium reaction is disadvantageous in that the metallic CNTs are covalently functionalized, the attachment of aryl groups to nanotubes has been found to be mostly reversible.11,12 However, this perturbation of the conjugated conductive π structure may also be used to “switch oﬀ” metallic nanotubes in thin ﬁlms or transistor devices without physically separating the semiconducting and metallic tubes.23,24 For any such deactivation or separation the overall eﬃcacy is determined by the initial selectivity of the functionalization process, as semiconducting nanotubes are always functionalized to some extent during such reactions.21 It is of interest to determine experimental conditions for which the reaction selectivity is maximized before separation techniques are employed. However, this is not a trivial exercise as the interaction of the diazonium species with the nanotube surface is dependent not only on the nanotube structure but also on the type of diazonium salt, its concentration, reaction temperature, solution pH, the nature of the surfactant, and the presence or absence of light.13,2022 In this report, arc-discharge produced CNTs are suspended in aqueous solution using a number of diﬀerent ionic and nonionic dispersants. The reactions of 4-nitro-, 4-bromo-, and 4-carboxy aryl diazonium salts with the nanotubes are analyzed in each case to examine the eﬀect of the type of dispersing medium on the metallic and semiconducting rates of reaction. The eﬀect of each dispersant on the dediazoniation process in the absence of nanotubes was also examined, as this was expected to aﬀect the reaction rate of the diazonium compounds with CNTs and possibly the selectivity.
2. EXPERIMENTAL METHODS Surfactants Brij S-100, Pluronic F-127 (PF-127), polyvinylpyrrolidone (PVP, MW ∼ 55 000), sodium dodecylbenzenesulfonate (SDBS), sodium dodecyl sulfonate (SDS), sodium deoxycholate (DOC), Triton X-405, sodium carboxymethylcellulose (SCMC, MW ∼ 90 000), Tween 60, and 4-aminobenzoic acid as well as 4-nitrobenzenediazonium tetraﬂuoroborate (nitroBDTFB) and 4-bromobenzenediazonium tetraﬂuoroborate (bromo-BDTFB) salts were purchased from Sigma-Aldrich (Sydney, Australia) and used as received. Hexadecyltrimethylammonium bromide (or cetyltrimethylammonium bromide, CTAB) was obtained from Ajax (Sydney, Australia).
4-Carboxybenzenediazonium tetraﬂuoroborate (carboxy-BDTFB) was synthesized according to the general procedure described by Roe.25 Speciﬁcally, 4-aminobenzoic acid (13.72 g, 0.1 mol) was dissolved in a mixture of 48% ﬂuoroboric acid (34 mL) and distilled water (40 mL). After cooling to 0 °C, sodium nitrite (6.8 g, 0.1 mol) in distilled water (15 mL) was added dropwise. The mixture was stirred for an additional 30 min, and the thick precipitate was collected by vacuum ﬁltration. The powder was puriﬁed by dissolving in a minimum amount of acetone and then ﬂocculated by addition of diethyl ether. The product 4-carboxy-BDTFB was obtained as a pale orange powder (20.00 g, 85%). Aqueous surfactant solutions were prepared at or slightly below their optimal concentrations (as per previously reported values)26 or at the standard 1 wt % where this value was not determined. Speciﬁcally, these concentrations were 0.5 wt % for SDBS, 4.5 wt % for Pluronic F-127, 3 wt % for PVP, 1.6 wt % for DOC, 3 wt % for TX-405, 2 wt % for Brij S-100 and 1 wt % for CTAB, SDS, SCMC, and Tween-60. As-produced electric arc (Carbon Solutions, Riverside, CA) or HiPCO (Carbon Nanotechnologies Inc., Houston, TX) CNTs were prepared according to a procedure optimized for dispersion in SDBS,27 wherein CNTs were added to 25 mL of aqueous surfactant solutions at a concentration of 0.5 mg mL1 via 30 min of tip ultrasonication using a Sonics VCX 750 W (Sonics, Newtown, CT) sonicator operating at 20 kHz with a 6.5 mm Ti microtip set to 20% of the maximum amplitude (corresponding to a power input of ∼0.28 W mL1). Samples were cooled with ice water during exposure to ultrasound and were ultracentrifuged directly after sonication for 1 h at ﬁxed angle at 37 000 rpm (∼1 105g) in a Type 70 Ti rotor mounted in an Optima L-100XP ultracentrifuge (Beckman-Coulter, Sydney, Australia), with the upper ∼7085% of the supernatant collected via pipet. As the amount of CNTs retained in the supernatant is variable among dispersants, collected volumes were further diluted with corresponding surfactant or polymer solutions to obtain roughly the same absorbance (0.55) at ∼850 nm and hence approximately the same concentration of CNTs (∼0.024 mg mL1) before addition of the diazonium reagent. In each case a fresh diazonium solution was prepared by dissolving a predetermined amount of the salt (22.0 mg of carboxy-, 22.1 mg of nitro-, and 25.2 mg of bromo-BDTFB) in 1.5 mL water through agitation and brief bath sonication (